With bellwether semiconductor company Nvidia set to present earnings after the markets close on Wednesday, all of Wall Street is anxious to see if the AI bull market still has legs.
Although founder Jensen Huang will feel the heat if he can’t deliver a quick quarterly report, the pressure this time isn’t as intense as it was in August, when it seemed the entire stock market’s fortunes turned after several big-name tech companies took a hit. was hanging on every word. At the time, Wedbush Securities’ Dan Ives called Nvidia’s results “the most important tech earnings in years.”
Still, Huang is expected to set a course for the industry moving forward as investors are eager for clues about the health of the artificial intelligence boom. Nvidia has been the biggest beneficiary of the investment wave, as it supplies nearly 9 out of 10 AI training chips to data centers.
Its shares have tripled in value since the start of this year, while the tech-heavy Nasdaq Composite has risen by just a third over the same period.
“We expect another jaw-dropper tomorrow from the godfather of AI Jensen that will inject jet fuel into this bull market engine,” Ives wrote Tuesday, confirming his fundamental investment thesis. In their view, Nvidia’s barely proof market share effectively means it is the “only game in town” and can expect $1 trillion or more in capex spending from consumers.
That doesn’t mean the bar isn’t rising, especially as Nvidia continues to lap up its easily comparable results from last year’s quarters before the GenAI boom.
Concern over rapid decline in GenAI investment
There is a growing debate surrounding the question of whether progress in neural networks is beginning to plateau, with the pace of innovation slowing, apparently due to the large amount of fresh data required to train large language models. ending
Meta’s chief AI scientist Yann Likon, one of the field’s original luminaries, cautioned that simply trying to scale models with more powerful chips won’t be enough. And can crush even more data. The approach will be enough.
“LLMs *will* not reach human-level intelligence,” he posted on Threads last week. “New architecture is needed.”
OpenAI’s recent product launch cadence is often cited as an example. Nearly two years since its predecessor’s launch in early 2023, GPT-5 is still nowhere to be seen. Meanwhile, initial plans to commercially launch Sora, its text-to-video GenAI tool, do not appear realistic before the end of the year, especially now that chief technology officer Meera Murthy has left the company.
If Nvidia customers see a diminishing return on their investment and come to a similar conclusion, they are likely to be more cautious with their orders moving forward.
Blackwell chips are getting hotter.
Secondly, Nvidia is facing a lot of questions surrounding its latest-generation AI training chip, dubbed Blackwell, which is effectively two AI chips sandwiched into one.
For example, Japan’s decision to make SoftBank its first Blackwell customer raised flags. A more obvious candidate would have been OpenAI, which in April received the first Hopper H200 delivered by Huang himself.
This unusual choice may be in part because early reports of design problems are holding back the rollout. Publication of technical news Information revealed that Blackwell chips are overheating in server racks that require particularly high power consumption of about 120 kW.
Nvidia confirmed that their installation required further adjustments based on this, calling it “normal and expected”.
However, analysts are not completely satisfied with this. “Investors now need to add this to the list of questions,” Wells Fargo said in a research note published on Sunday.
Importantly, Blackwell is not expected to influence the fiscal third-quarter results that Nvidia will report on Wednesday, but it could affect their outlook.
That’s because Huang has promised that the company will already generate several billion dollars in revenue in the current fourth quarter, which ends in late January.
Finally another concern this time — completely out of Huang’s control — is the accounting issues behind Super Microcomputer, which purchases Nvidia AI chips as part of its data center hardware business.
Just earlier this year, a recent report from investor darling, short-seller Hindenburg Research raised fundamental questions about the company’s accounting practices. Last month its auditor completely jumped ship. Super Micro announced on Monday that it has hired BDO USA as its new auditor.
Mark Yusko, CEO and head of investments at Morgan Creek Capital, with $1.9 billion in assets under management, said when asked about Nvidia’s current valuation given the problems surrounding SuperMicro CEO Charles Liang’s firm. When they went, they were scared.
“What happens to your revenue growth and profit growth if they don’t have the ability to be your third largest customer? We’ve seen that before with Nortel and Cisco in 2000 and 2001,” Mark Yusko told CNBC last week. “I’ll be the seller.”
About half of Nvidia’s revenue comes from just four major customers, the company revealed in August, raising the risk of concentration for the chip supplier.
